Background: The clinical dementia rating (CDR) scale is commonly used to diagnose dementia due to Alzheimer’s disease (AD). The sum of boxes of the CDR (CDR-SB) has recently been emphasized and applied to interventional trials for tracing the progression of cognitive impairment (CI) in the early stages of AD. We aimed to study the influence of baseline CDR-SB on disease progression to dementia or reversion to normal cognition (NC).Materials and methods: The baseline CDR < 1 cohort registered from September 2015 to August 2020 with longitudinal follow-up in the History-based Artificial Intelligence Clinical Dementia Diagnostic System (HAICDDS) database was retrospectively analyzed for the rates of conversion to CDR ≥ 1. A Cox regression model was applied to study the influence of CDR-SB levels on progression, adjusting for age, education, sex, neuropsychological tests, neuropsychiatric symptoms, parkinsonism, and multiple vascular risk factors.Results: A total of 1,827 participants were analyzed, including 1,258 (68.9%) non-converters, and 569 (31.1%) converters with mean follow-up of 2.1 (range 0.4–5.5) and 1.8 (range 0.3–5.0) years, respectively. Conversion rates increased with increasing CDR-SB scores. Compared to a CDR-SB score of 0, the hazard ratios (HR) for conversion to dementia were 1.51, 1.91, 2.58, 2.13, 3.46, 3.85, 3.19, 5.12, and 5.22 for CDR-SB scores of 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, and ≥4.5, respectively (all p < 0.05 except for CDR-SB score = 0.5). In addition, older age, lower education, lower cognitive performance, and a history of diabetes also increased conversion rates. Furthermore, reversions to NC were 12.5, 5.6, 0.9, and 0% for CDR-SB scores of 0.5, 1.0–2.0, 2.5–3.5 and ≥4.0, respectively (p < 0.001).Conclusion: CDR-SB in predementia or very mild dementia (VMD) stages highly predicts progression to dementia or reversion to NC. Therefore, CDR-SB could be a good candidate for tracing the effectiveness of pharmacological and non-pharmacological interventions in populations without dementia.

Objective. The reliable diagnosis remains a challenging issue in the early stages of dementia. We aimed to develop and validate a new method based on machine learning to help the preliminary diagnosis of normal, mild cognitive impairment (MCI), very mild dementia (VMD), and dementia using an informant-based questionnaire. Methods. We enrolled 5,272 individuals who filled out a 37-item questionnaire. In order to select the most important features, three different techniques of feature selection were tested. Then, the top features combined with six classification algorithms were used to develop the diagnostic models. Results. Information Gain was the most effective among the three feature selection methods. The Naive Bayes algorithm performed the best (accuracy = 0.81, precision = 0.82, recall = 0.81, and F-measure = 0.81) among the six classification models. Conclusion. The diagnostic model proposed in this paper provides a powerful tool for clinicians to diagnose the early stages of dementia.

Objectives: The Clinical Dementia Rating (CDR) Scale is the gold standard for the staging of dementia due to Alzheimer's disease (AD). However, the application of CDR for the staging of subjective cognitive decline (SCD) and mild cognitive impairment (MCI) in AD remains controversial. This study aimed to use the sum of boxes of the CDR (CDR-SB) plus an SCD single questionnaire to operationally determine the different stages of cognitive impairment (CI) due to AD and non-AD.Methods: This was a two-phase study, and we retrospectively analyzed the Show Chwan Dementia registry database using the data selected from 2015 to 2020. Individuals with normal cognition (NC), SCD, MCI, and mild dementia (MD) due to AD or non-AD with a CDR < 2 were included in the analysis.Results: A total of 6,946 individuals were studied, including 875, 1,009, 1,585, and 3,447 with NC, SCD, MCI, and MD, respectively. The cutoff scores of CDR-SB for NC/SCD, SCD/MCI, and MCI/dementia were 0/0.5, 0.5/1.0, and 2.5/3.0, respectively. The receiver operating characteristic (ROC) analysis showed that the area under the curve (AUC) values of the test groups were 0.85, 0.90, and 0.92 for discriminating NC from SCD, SCD from MCI, and MCI from dementia, respectively. Compared with the Cognitive Abilities Screening Instrument or the Montreal Cognitive Assessment, the use of CDR-SB is less influenced by age and education.Conclusion: Our study showed that the operational determination of SCD, MCI, and dementia using the CDR-SB is practical and can be applied in clinical settings and research on CI or dementia.

Objectives: The association of vascular risk factors (VRFs) with incidence of Alzheimer's disease (AD) and vascular dementia (VaD) has been well studied. However, the association between VRFs and non‐AD dementia is seldom investigated. In this study, we aim to compare the concurrence of VRFs of Lewy body dementia (LBD) to AD.Materials & Methods: We consecutively enrolled patients with dementia with Lewy bodies (DLB), Parkinson's disease dementia (PDD), and AD, and the prevalence of arterial hypertension, diabetes mellitus, hypercholesterolemia, hyperlipidemia, smoking, and obesity was assessed and compared.Results: A total of 167 consecutive patients were studied. Among them, 55 had DLB, 34 had PDD, and 78 had AD. History of any VRF among three groups was not significantly different. In addition, the patients with AD had significantly higher systolic pressure (SBP), diastolic pressure (DBP), waist, body mass index (BMI), ac glucose, and HbA1c (all p‐value < .005). After a stepwise procedure that considered age, sex, disease severity, antiparkinson drugs, systolic and diastolic blood pressures, glycated hemoglobin, body mass index, triglyceride, total cholesterol, and low‐density lipoprotein, SBP in the third tertile (144–198 mmHg), BMI in the second tertile (23.6–26.4), and TG in the third tertile (127–310) were significantly lower in LBD than in AD.Conclusions: The VRFs are less associated with LBD than with AD. DLB and PDD showed a similar pattern of the association of VRFs. SBP, BMI, and TG were significantly lower in LBD compared with AD.

Objective: This paper proposes a multiclass deep learning method for the classification of dementia using an informant-based questionnaire. Methods: A deep neural network classification model based on Keras framework is proposed in this paper. To evaluate the advantages of our proposed method, we compared the performance of our model with industry-standard machine learning approaches. We enrolled 6,701 individuals, which were randomly divided into training data sets (6030 participants) and test data sets (671 participants). We evaluated each diagnostic model in the test set using accuracy, precision, recall, and F1-Score. Results: Compared with the seven conventional machine learning algorithms, the DNN showed higher stability and achieved the best accuracy with 0.88, which also showed good results for identifying normal (F1-score=0.88), mild cognitive impairment (MCI) (F1-score=0.87), very mild dementia (VMD) (F1-score=0.77) and Severe dementia (F1-score=0.94). Conclusion: The deep neural network (DNN) classification model can effectively help doctors accurately screen patients who have normal cognitive function, mild cognitive impairment (MCI), very mild dementia (VMD), mild dementia (Mild), moderate dementia (Moderate), and severe dementia (Severe).